Intersections of political and economic relations: a network study

Transcription

1 Procedia Computer Science Volume 66, 2015, Pages YSC th International Young Scientists Conference on Computational Science Intersections of political and economic relations: a network study Zsolt Főző 1,László Gulyás 2, and George Kampis 3,4 1 Rajk László College for Advanced Studies, Budapest zsolt.fozo@gmail.com 2 ELTE University, Budapest gulyas@hps.elte.hu 3 ELTE University, Budapest gk@hps.elte.hu 4 ITMO University, St. Petersburg, Russia Abstract In this paper we analyze a special subset of the Hungarian aháló network database, one that contains different economic, political and personal relations in a form of a single network. We narrowed down the original dataset to specific types of economic, political and personal relations. The sampled network has small-world, scale-free and non-transitive properties. Our hypothesis was that many actors (i.e. nodes in the network) share different types of economic and political relationships. This hypothesis is not confirmed in the study, as we only found a low number of node-pairs that share a given type of economic link and some type of political link at a time. Keywords: social network analysis, computational social science, politics and economy, corruption studies 1 Introduction It is of high importance to study the relationship between economic and political spheres. These relations can reveal possible corruption, can serve as attributes of society s institutions, can reflect politicalbusiness cycles, or reveal the average atmosphere and level of trust in a country. In this paper we analyze such relations on a Hungarian database that contains several connections between the actors of economic and political life. Since the numeric value of the country s Corruption Perception Index is known to be quite high (it has a value of 53 on a 100 point scale - according to Transparency International [1]) we expect valuable observations to emerge where the political and economic relations are thus merged. The current work also merits attention because it is one of the first analyses of this major Hungarian social and political network database called aháló. Here we first sketch the basic aspects of the analysis of political-economic relations in general, then introduce and analyze the particular network under study, and test our a priori expectation that personal, Selection and peer-review under responsibility of the Scientific Programme Committee of YSC 2015 c The Authors. Published by Elsevier B.V. doi: /j.procs

2 political and economic dimensions are strongly connected in the available data. 2 Relations between the economic and political spheres In this section we introduce the studies that emphasize the importance of politicians business connections (and the other way around). In this short paper we do not have the possibility to explore and analyze truly complex relationships, yet it may important to illustrate the basics. One of the first questions that can be raised is how companies and actors that possess direct political connections perform in the market. Are they more successful economically? Amy J. Hillman [8] tried to answer this in her article Politicians on the board of directors. The author analyzed companies in both heavily regulated markets and less regulated markets, and searched for how common is that (ex- )politicians are at the same (or a different) time high level executives of the companies. She found that politicians are more typical economic actors in those markets that are more heavily regulated, moreover, that companies led by politicians are preforming better in these markets. Bertrand et al. [2] have studied similarly the performance of well-connected companies in the light of political-business cycles. In their research, they analyzed what happened with stock-exchange companies that feature ex-politician executives in the years of national elections. Their results show that in the crucial years of elections, these companies were more active in employing new people and extending capacities than their not-so-well-connected counterparts, even though state support or trade aid was not effecting these processes directly. It is worth highlighting that at the same time, because of their raising costs (salaries and costs of capacity extensions), these companies realized lower profit levels. Schleifer and Vishny [10] modeled the relationships of companies and politicians. In their article Politicians and firms they outlined a model, where, if politicians are leading companies (such as public firms or freshly privatized former state firms) the managers follow a set of practices that are independent from the political goals, whereas, if genuine managers are leading the company, the politicians try to bribe them with concessions, aid or money to make them act more politically. According to their model, such rent-type concessions and aids are more harmful in the case of politician-led companies; therefore, they conclude that successful privatization can lead to a more efficient market equilibrium [10]. This model influenced other researchers as well: Claessens and Djankov [6] studied the same effects in the Central-East European environment after the collapse of Communism and in the resulting transformation. They find that an effective privatization process can indeed lead to more profitable companies and more effective market equilibrium - also in the case of not politically led companies. Bunkanwanicha and Wiwattanakantang [5] studied other aspects of business-politics relations in their article emph Big business owners in politics. Using a Thailand database, they state that the wealthier the leader of a company, the more plausible is that she/he will run for high political positions. As politicians, then, these people are more likely to support policies in the interests of their own company, although they are typically not supporting their own companies with public resources directly. Let us summarize the results. These studies show various effects of political influences in the market environment. In the short term, political influence can boost the performance of companies in those sectors, where keeping good relationship with the state is important. On the contrary, if a company operation serves direct political goals, this may cause a worse performance with fewer profits. Such a political goal can include labor and capacity expansion according to political-business cycles that may lead to less effective operations and less profits. Moreover, it is worth mentioning that not only politicians may have business interests, but also business leaders may try to run for political positions. These political positions help them influence state regulation and the economic environment, but do not lead to a direct financial support of the companies from public sources. 240

3 3 The database 3.1 Background: the K-Monitor organization For the following analyses we use the aháló database of the K-Monitor civil organization from Hungary. The K-Monitor Independent Public Financial Monitoring Service is a civil organization that fights against the spread of corruption in the country by collecting data from public procurement procedures. This civil organization collects data about public spending and corruption from the public media, and organizes them into a library. At the website of K-Monitor 1, we can find neutral, aggregated information about actors and activities, helping the creation of public discourses about them. The organization continuously develops the database using automatized search and parsing as well as manual work. 3.2 The aháló project K-Monitor operates the aháló project, producing a database developed under the supervision of the second and third authors. The goal of the project is to automatically collect and aggregate public data about the network of Hungarian politicians and business actors. The data is made available in the form of a visualized network at the project website 2 that helps achieve understanding and transparency. The aháló database currently contains information from: Procurement reports, The database of National Development Agency (state support), The data of Bureau for agriculture and rural development (state support) List of state executives (since 1990) List of municipal officials List of companies owned be politicians and party members List companies that had public mandate Newspaper articles about business relations and about the use of public sources. This major database has hardly been analyzed yet. The purpose of the present paper is to describe the characteristics of the database with repsect to the different type of relations, highlighting those that represent both political and economic ties between actors. 4 Analyses of networks built from the database 4.1 The entire network of aháló The whole database can be handled as a single big network, where there are currently 169,459 nodes (i.e. persons and organizations) and 1,748,040 directed links [7]. There are three major categories of links: links between persons (P2P), links between organizations (O2O) and links between persons and organizations (P2O). The different types of relations can be attributed with different edges that arise from the various data sources. There are altogether 140 types of detailed connections in the database

4 1 P2P-advisor 11 P2O-owner 2 P2P-friend 12 P2O- Company partner (full) 3 P2P-fellow representatives 13 P2O-Company partner (silent) 4 P2P-common organizational background 14 P2O-Founder/shareholder 5 P2P-relative 15 P2O- Founder/shareholder (1 person) 6 P2P-fellow municipality representative 16 P2O-Member of general partnership 7 P2P-candidate of the same party 17 P2O-ineterest/subsidiary (Ltd) 8 P2P-fellow candidate 18 P2O-Private company owner 9 P2P-fellow representatives in the Parliament 19 P2O-member of an association 10 P2P-fellow representatives in the EP Table 1: Types of edges in the reduced network (Source: selection from the aháló database) 4.2 The narrower network of personal relations As we are interested in the economic relations of politicians, we analyze P2P and P2O links only. According to this criterion, we have narrowed down the database to just 19 types of edges (Table 1). The reduced network has 35,393 nodes and 387,210 edges. 4.3 Properties of the reduced network We first examine those main properties of the network that generally serve as useful information: is it a small-world network, is it scale-free, is it clustered? [9]. Figure 1: Degree distribution of the reduced network. Left: in-degree, rigtht: out-degree Small-world property. Small world networks are those where the typical distance between two randomly chosen nodes grows proportionally to the logarithm of the number of nodes in the network [9]. In the case of our network, this value is 4,248, and the network is clearly a small-world network. Scale-free property. A scale-free network is one where the degree distribution follows a power law [9]. This practically means that besides many low degree edges, there are a few very high-degree edges (that are connected to many other edges, in turn). In our case, the network approximates a scale-free distribution: while the average degree is 10.49, there are a few edges connected to more than 20,000 other edges (both in- and out degrees). Figure 1 shows a log-log plot for visual inspection. 242

5 Clustering coefficient (i.e. transitivity property). The clustering coefficient C of a network shows how well the neighboring edges connect to each other ( my-friends-are-friends property). More precisely, C is the ratio of links connecting a node s neighbors divided by the maximum possible number of such links. The property thus measured can also be called transitivity for the obvious reason [9]. In our network, C is very low (0.1), so we can conclude that the network is non-transitive. 4.4 The intersection of different types of edges As seen in Table 1, different types of P2P and P2O correspond to different types of edges. We could thus construct several sub-networks for all different types of edges, which would show a more nuanced view of each specific type of relationship. The different types of edges can be sorted into three groups: Political relationship (3, 6, 7, 8, 9, 10) Economic relationship (11-19) Personal relationship (1, 2, 5) One of the main goals of this paper is to analyze how political relationships share the links of economic relationships. Thus we look for cases where persons (i.e. nodes) can be connected with more than one types of relationships. In order to find these cases, we collect the number of nodes that can be affiliated with different types of relationships at the same time. This serves as a test of our hypothesis that such relations are strongly manifest in the dataset and that personal political and economic ties are closely related in many cases. Below (Table 2) we show a matrix of tables, where different vectors made from a series of common nodes are shown in the context of different political and economic relationships. We highlight the numbers of interest. In these tables, we can see that relative to the aforementioned high number of edges and links, there are just a very few nodes with more than one different types of connections with another node. And the most common of them is the one when candidates of the same party are also company partners (135 full partners, 156 silent partners), and this is the most frequent type of political link in the network. As we possess data for the purely personal relations as well, we also have the possibility to study the intersection of personal and economic relations. In Table 3 we can see that there are just a few nodes that share the same economic and personal relations. Again, company members can be found among friends and relatives, but these are extremely low numbers, if we think about the size of the network or the number of the different types of links. So in conclusion we cannot support the fundamental hypothesis that - at the level of available data - personal political and economic ties would strongly intertwined. Seeking for the reasons leads us to our discussion section. 5 Discussion In this short paper we have analyzed a special subset of the aháló network database. We narrowed down the original dataset to different types of economic, political and personal relations. The resulting network is a small-world, essentially scale-free and non-transitive network. (Other approaches to key network properties have not been considered in this paper, such as those based on link prediction [3] or significance [4]. Similar studies are left to future work.) Our initial hypothesis was that many actors (i.e. nodes in the network) share different types of economic and political relationships. This hypothesis does not seem to hold, however, since we only found a relatively low number node-pairs that share some type of economic link and some type of 243

6 Fellow representatives Fellow representatives at municipality level Owner 1 15 Company partner (full) 0 70 Company partner (silent) 0 93 Funder/shareholder 0 18 Funder/shareholder ( person) Member of a gen. partnership 0 1 Interest/subsidiary (Ltd) 0 0 Private company owner 0 0 Member of an association 0 0 Fellow party candidates Fellow candidate Owner 32 0 Company partner (full) Company partner (silent) Funder/shareholder 35 0 Funder/shareholder ( person) Member of a general partnership 7 0 Interest/subsidiary (Ltd) 0 0 Private company owner 2 0 Member of an association 0 0 Fellow representatives in the Parliament Fellow representatives in the EP Owner 7 0 Company partner (full) 36 1 Company partner (silent) 47 7 Funder/shareholder 2 0 Funder/shareholder ( person) Member of a general partnership 1 0 Interest/subsidiary (Ltd) 0 0 Private company owner 0 0 Member of an association 0 0 Table 2: Matrices of different political and economic relationships 244

7 Advisor Friend Relative Owner Company partner (full) Company partner (silent) Funder/shareholder Funder/shareholder (1 person) Member of a general partnership Interest/subsidiary (Ltd) Private company owner Member of an association Table 3: Matrix of personal and economic relationships political link simultaneously. Yet we nevertheless also see that some interlinks between politics and economy exist - to be examined in future work. On the other hand, we have to carefully consider the interpretation of the findings because of the limitations of the research. One of the biggest challenges is the quality of data. Analyzing connections between people is a new field of science and as such researches lack experience of data collection. As connections can vary on a wide scale, it is hard to define what kind of data should be collected. For the current research the biggest obstacle is that connections are unweighted, and as a consequence, the intensity of business connections are not measurable (such as the volume and value of business transactions). Analytically speaking, we can identify the following limitations or rather specifics of the data collection process: 1. Too little data, resulting in homogenous business connections. Business partnerships between small SMEs are just treated the same way as billion dollar transactions between large business partners. As the edges only able to show connections between people, the findings regarding transfers, wealth or financial well-being would thus be unreliable. Under the current circumstances the intensity of connections solely depends on the number of business connections between partners in the database. 2. The research involves business connections only. As such it is unable to reflect other connection types that may affect business connections (think of connections such as friendships among influential individuals). Such a limitation can restrict the findings in case of Hungary. As a result of increasing volume of public investments in all economic sectors and high level of corruption businessmen having rather political than business connections can easily gain power within the network of economic and political elite. 3. Lack of proper data. Referring to the second section, it is important to note that (1) it is not evident what type of data should be recorded and (2) all the data is from publicly available sources that means the analyzed network do not involve not public connections between people. As a result there may be connections between people that we do not know about. 6 Acknowledgements Collaboration with K-Monitor, Ltd. and Sándor Lederer are gratefully acknowledged. Data is owned by K-Monitor and used here with their permission. This work was partially supported by the Russian 245

8 Scientific Foundation, proposal # , Supercomputer simulation of critical phenomena of complex social systems. References [1] Corruption Perception Index Ranking. Transparency International, [2] M. Bertrand, F. Kramarz, A. Schoar, and D. Thesmar. Politicians, firms and the political business cycle: evidence from France. Unpublished working paper. Chicago: University of Chicago, [3] C. A. Bliss, M. R. Frank, C. M. Danforth, and P. S. Dodds. An evolutionary algorithm approach to link prediction in dynamic social networks. Journal of Computational Science, 5(5): , [4] I. Blokh and V. Alexandrov. Discovering most significant news using network science approach. Procedia Computer Science, 51: , [5] P. Bunkanwanicha and Y. Wiwattanakantang. Big business owners in politics. Review of financial studies, 22(6): , [6] S. Claessens and S. Djankov. Politicians and firms in seven Central and Eastern European countries. World Bank Policy Research Working Paper, [7] L. Gulyás and G. Kampis. NetSCi overview of the aháló Network. unpublished manuscript, [8] A. J. Hillman. Politicians on the board of directors: Do connections affect the bottom line? Journal of Management, 31(3): , [9] M. Newman. Networks: An introduction. Oxford: Oxford University Press, [10] A. Shleifer and R. W. Vishny. Politicians and firms. The Quarterly Journal of Economics, pages ,